[Amps] low Q output matching

[Manfred Mornhinweg]

Carel,

> I don't mind
> even 0 dB of harmonic suppression.... because that is easily solved
> afterwards with fixed valued C's/L's lowpass filter. 

If you have no harmonic suppression in the matching circuit, then you 
will probably need a 7 pole low pass filter, at least, after a single 
ended class AB amplifier. That starts to become complex and expensive, 
specially once you realize how hard to find and expensive the capacitors 
for such a filter are!

Or if you know where to get suitable capacitors cheaply, please let me 
know, because I'm looking for some, and the prices I have seen tell me 
very clearly that I should better fabricate my own capacitors!

The problem is that none of the commonly available inexpensive ceramic 
or foil capacitors can handle the RF current present in such a filter. 
You eithe rhave to use the best of them, overload them to three or five 
times their rating and hope for the best, like certain manufacturers are 
doing, or you have to buy some of the very select and expensive high 
current capacitors specially made for RF power.

Now when you need 24 of those capacitors to make a bank of six 
switchable 7 pole filters, the cost can be quite intimidating.

> In essence, could I just
> a high power 3kW broadband 50 to 1700 Ohm RF transformer and the tube would
> be happy ?

No, unfortunately you cannot. Absolutely not!

I asked myself the same question some time ago. My specific question 
was: Why has nobody ever come out with a broadband amp and switchabele 
low pass filters, using the same tubes we normally use?

The answer is the same I mentioned about low Q tuned circuits in my mail 
of yesterday: The huge tube output capacitance, which you MUST absorb 
somewhere!

For example, take the old 813 tube. It has 14pF output capacitance, 
each. Four of them in parallel would give a load impedance that would 
suit your 1700 to 50 Ohm broadband transformer, if you choose the right 
plate voltage. But those four 813's would have 56pF output capacitance! 
Add some stray capacitance, and you might have 70pF total. On the 10 
meter band that's a reactance of only 78 Ohm!!! This will totally short 
out the tube's output signal! There is absolutely no way to make this work.

So you need to tune out those 70pF, for example by replacing your high 
impedance plate choke with a low inductance coil that has 78 ohm 
inductive reactance. This would form a parallel resonant circuit with 
the tube's and stray capacitance. But that, in combination with your 
perfect, zero-Q broadband transformer (if you can make such a beast - I 
can't!), still gives a Q of 22! And that inductor would need to 
bandswitched and tuned, of course.

With more modern tubes, rated for higher frequency, that have lower 
capacitances relative to their optimal load resistance, the situation is 
quite a bit better, but never good enough to run a big tube on 10 meters 
in a broadband circuit.

That's why tubes at RF are always used with tuned circuits, and often 
with a rather high Q.

Regarding flywheel effect and restitution of the half cycle in which the 
tube doesn't conduct, that's the exact same thing as suppressing the 
harmonics. Any low pass filter that achieves good harmonic suppression 
will nicely restore the sine waveshape. But no low-Q low pass filter can 
tune out the tube's capacitance.

Manfred

========================
Visit my hobby homepage!
http://ludens.cl
========================